Machine‐learning‐based predictive control of nonlinear processes. Part II: Computational implementation

Wu, Zhe; Tran, Anh; Rincon, David; Christofides, Panagiotis D.

doi:10.1002/aic.16734

Title: Machine‐learning‐based predictive control of nonlinear processes. Part II: Computational implementation

Abstract

Abstract Machine learning is receiving more attention in classical engineering fields, and in particular, recurrent neural networks (RNNs) coupled with ensemble regression tools have demonstrated the capability of modeling nonlinear dynamic processes. In Part I of this two‐article series, the Lyapunov‐based model predictive control (LMPC) method using a single RNN model and an ensemble of RNN models, respectively, was rigorously developed for a general class of nonlinear systems. In the present article, computational implementation issues of this new control method ranging from training of the RNN models, ensemble regression of the RNN models, and parallel computation for accelerating the real‐time model calculations are addressed. Furthermore, a chemical reactor example is used to demonstrate the implementation and effectiveness of these machine‐learning tools in LMPC as well as compare them with standard state‐space model identification tools.

Authors:

Wu, Zhe ^[1]; Tran, Anh ^[1]; Rincon, David ^[1];

^[2]

Department of Chemical and Biomolecular Engineering University of California Los Angeles California
Department of Chemical and Biomolecular Engineering University of California Los Angeles California, Department of Electrical and Computer Engineering University of California Los Angeles California

Publication Date:: 2019-07-31

Sponsoring Org.:: USDOE

OSTI Identifier:: 1545905

Resource Type:: Publisher's Accepted Manuscript

Journal Name:: AIChE Journal

Additional Journal Information:: Journal Name: AIChE Journal Journal Volume: 65 Journal Issue: 11; Journal ID: ISSN 0001-1541

Publisher:: Wiley Blackwell (John Wiley & Sons)

Country of Publication:: United States

Language:: English

Citation Formats


                    Wu, Zhe, Tran, Anh, Rincon, David, and Christofides, Panagiotis D. Machine‐learning‐based predictive control of nonlinear processes. Part II: Computational implementation.  United States: N. p., 2019. 
Web.  doi:10.1002/aic.16734.

Copy to clipboard


                    Wu, Zhe, Tran, Anh, Rincon, David, & Christofides, Panagiotis D. Machine‐learning‐based predictive control of nonlinear processes. Part II: Computational implementation.  United States.  https://doi.org/10.1002/aic.16734

Copy to clipboard


                    Wu, Zhe, Tran, Anh, Rincon, David, and Christofides, Panagiotis D. Wed .  
"Machine‐learning‐based predictive control of nonlinear processes. Part II: Computational implementation".  United States.  https://doi.org/10.1002/aic.16734.

Copy to clipboard


                    
@article{osti_1545905,

  title        = {Machine‐learning‐based predictive control of nonlinear processes. Part II: Computational implementation},

  author       = {Wu, Zhe and Tran, Anh and Rincon, David and Christofides, Panagiotis D.},

  abstractNote = {Abstract Machine learning is receiving more attention in classical engineering fields, and in particular, recurrent neural networks (RNNs) coupled with ensemble regression tools have demonstrated the capability of modeling nonlinear dynamic processes. In Part I of this two‐article series, the Lyapunov‐based model predictive control (LMPC) method using a single RNN model and an ensemble of RNN models, respectively, was rigorously developed for a general class of nonlinear systems. In the present article, computational implementation issues of this new control method ranging from training of the RNN models, ensemble regression of the RNN models, and parallel computation for accelerating the real‐time model calculations are addressed. Furthermore, a chemical reactor example is used to demonstrate the implementation and effectiveness of these machine‐learning tools in LMPC as well as compare them with standard state‐space model identification tools.},

  doi          = {10.1002/aic.16734},

  journal      = {AIChE Journal},

  number       = 11,

  volume       = 65,

  place        = {United States},

  year         = {2019},

  month        = {7}

}

Copy to clipboard

Journal Article:

Free Publicly Available Full Text

Accepted Manuscript (Publisher)

Publisher's Version of Record
https://doi.org/10.1002/aic.16734

Other availability

Search WorldCat to find libraries that may hold this journal

Citation Metrics:

Cited by: 60 works

Citation information provided by
Web of Science

Save / Share:

Export Metadata

Save to My Library

Works referenced in this record:

Constrained Nonlinear State Estimation Using Ensemble Kalman Filters
journal, March 2010

Prakash, J.; Patwardhan, Sachin C.; Shah, Sirish L.
Industrial & Engineering Chemistry Research, Vol. 49, Issue 5
DOI: 10.1021/ie900197s

A parallel computing platform for training large scale neural networks
conference, October 2013

Gu, Rong; Shen, Furao; Huang, Yihua
2013 IEEE International Conference on Big Data
DOI: 10.1109/BigData.2013.6691598

The Ensemble Kalman Filter: theoretical formulation and practical implementation
journal, November 2003

Evensen, Geir
Ocean Dynamics, Vol. 53, Issue 4
DOI: 10.1007/s10236-003-0036-9

Economic model predictive control of nonlinear process systems using empirical models
journal, November 2014

Alanqar, Anas; Ellis, Matthew; Christofides, Panagiotis D.
AIChE Journal, Vol. 61, Issue 3
DOI: 10.1002/aic.14683

David Draper and E. I. George, and a rejoinder by the authors
journal, November 1999

Volinsky, Chris T.; Raftery, Adrian E.; Madigan, David
Statistical Science, Vol. 14, Issue 4
DOI: 10.1214/ss/1009212519

Parallel distributed computing using Python
journal, September 2011

Dalcin, Lisandro D.; Paz, Rodrigo R.; Kler, Pablo A.
Advances in Water Resources, Vol. 34, Issue 9
DOI: 10.1016/j.advwatres.2011.04.013

The ensemble Kalman filter for combined state and parameter estimation
journal, June 2009

Evensen, Geir
IEEE Control Systems, Vol. 29, Issue 3
DOI: 10.1109/MCS.2009.932223

A universal formula for stabilization with bounded controls
journal, June 1991

Lin, Yuandan; Sontag, Eduardo D.
Systems & Control Letters, Vol. 16, Issue 6
DOI: 10.1016/0167-6911(91)90111-Q

Data Driven Economic Model Predictive Control
journal, April 2018

Kheradmandi, Masoud; Mhaskar, Prashant
Mathematics, Vol. 6, Issue 4
DOI: 10.3390/math6040051

Recurrent neural network for solving model predictive control problem in application of four-tank benchmark
journal, May 2016

Xu, Jing; Li, Chuandong; He, Xing
Neurocomputing, Vol. 190
DOI: 10.1016/j.neucom.2016.01.020

Model Predictive Control of Unknown Nonlinear Dynamical Systems Based on Recurrent Neural Networks
journal, August 2012

Pan, Yunpeng; Wang, Jun
IEEE Transactions on Industrial Electronics, Vol. 59, Issue 8
DOI: 10.1109/TIE.2011.2169636

An Ensemble Kalman Filter for Systems Governed by Differential Algebraic Equations (DAEs)
journal, January 2012

Puranik, Yash; Bavdekar, Vinay A.; Patwardhan, Sachin C.
IFAC Proceedings Volumes, Vol. 45, Issue 15
DOI: 10.3182/20120710-4-SG-2026.00167

On the implementation of an interior-point filter line-search algorithm for large-scale nonlinear programming
journal, April 2005

Wächter, Andreas; Biegler, Lorenz T.
Mathematical Programming, Vol. 106, Issue 1
DOI: 10.1007/s10107-004-0559-y

A framework for parallel and distributed training of neural networks
journal, July 2017

Scardapane, Simone; Di Lorenzo, Paolo
Neural Networks, Vol. 91
DOI: 10.1016/j.neunet.2017.04.004

Similar Records in DOE PAGES and OSTI.GOV collections:

Machine‐learning‐based construction of barrier functions and models for safe model predictive control

Journal Article Chen, Scarlett ; Wu, Zhe ; Christofides, Panagiotis D. - AIChE Journal

Abstract In this paper, we propose a control Lyapunov‐barrier function‐based model predictive control method utilizing a feed‐forward neural network specified control barrier function (CBF) and a recurrent neural network (RNN) predictive model to stabilize nonlinear processes with input constraints, and to guarantee that safety requirements are met for all times. The nonlinear system is first modeled using RNN techniques, and a CBF is characterized by constructing a feed‐forward neural network (FNN) model with unique structures and properties. The FNN model for the CBF is trained based on data samples collected from safe and unsafe operating regions, and the resulting FNNmore »« less
https://doi.org/10.1002/aic.17456
Using Machine Learning Tools to Predict Compressor Stall

Journal Article Hipple, Samuel M. ; Bonilla-Alvarado, Harry ; Pezzini, Paolo ; ... - Journal of Energy Resources Technology

Clean energy has become an increasingly important consideration in today’s power systems. As the push for clean energy continues, many coal-fired power plants are being decommissioned in favor of renewable power sources such as wind and solar. However, the intermittent nature of renewables means that dynamic load following traditional power systems is crucial to grid stability. With high flexibility and fast response at a wide range of operating conditions, gas turbine systems are poised to become the main load following component in the power grid. Yet, rapid changes in load can lead to fluid flow instabilities in gas turbine powermore »« less
https://doi.org/10.1115/1.4046458

Full Text Available
A cyber‐secure control‐detector architecture for nonlinear processes

Journal Article Chen, Scarlett ; Wu, Zhe ; Christofides, Panagiotis D. - AIChE Journal

Abstract This work presents a detector‐integrated two‐tier control architecture capable of identifying the presence of various types of cyber‐attacks, and ensuring closed‐loop system stability upon detection of the cyber‐attacks. Working with a general class of nonlinear systems, an upper‐tier Lyapunov‐based Model Predictive Controller (LMPC), using networked sensor measurements to improve closed‐loop performance, is coupled with lower‐tier cyber‐secure explicit feedback controllers to drive a nonlinear multivariable process to its steady state. Although the networked sensor measurements may be vulnerable to cyber‐attacks, the two‐tier control architecture ensures that the process will stay immune to destabilizing malicious cyber‐attacks. Data‐based attack detectors are developedmore »« less
Cited by 15
https://doi.org/10.1002/aic.16907
Accelerating phase-field predictions via recurrent neural networks learning the microstructure evolution in latent space

Journal Article Hu, C. ; Martin, S. ; Dingreville, R. - Computer Methods in Applied Mechanics and Engineering

The phase-field method is a popular modeling technique used to describe the dynamics of microstructures and their physical properties at the mesoscale. However, because in these simulations the microstructure is described by a system of continuous variables evolving both in space and time, phase-field models are computationally expensive. They require refined spatio-temporal discretization and a parallel computing approach to achieve a useful degree of accuracy. As an alternative, we present and discuss an accelerated phase-field approach which uses a recurrent neural network (RNN) to learn the microstructure evolution in latent space. We perform a comprehensive analysis of different dimensionality-reduction methodsmore »« less
https://doi.org/10.1016/j.cma.2022.115128
Robust and scalable uncertainty estimation with conformal prediction for machine-learned interatomic potentials

Journal Article Hu, Yuge ; Musielewicz, Joseph ; Ulissi, Zachary W. ; ... - Machine Learning: Science and Technology

Abstract Uncertainty quantification (UQ) is important to machine learning (ML) force fields to assess the level of confidence during prediction, as ML models are not inherently physical and can therefore yield catastrophically incorrect predictions. Established a-posteriori UQ methods, including ensemble methods, the dropout method, the delta method, and various heuristic distance metrics, have limitations such as being computationally challenging for large models due to model re-training. In addition, the uncertainty estimates are often not rigorously calibrated. In this work, we propose combining the distribution-free UQ method, known as conformal prediction (CP), with the distances in the neural network’s latent spacemore »« less
https://doi.org/10.1088/2632-2153/aca7b1

Similar Records

Title: Machine‐learning‐based predictive control of nonlinear processes. Part II: Computational implementation

Abstract

Citation Formats

Constrained Nonlinear State Estimation Using Ensemble Kalman Filters journal, March 2010

A parallel computing platform for training large scale neural networks conference, October 2013

The Ensemble Kalman Filter: theoretical formulation and practical implementation journal, November 2003

Economic model predictive control of nonlinear process systems using empirical models journal, November 2014

David Draper and E. I. George, and a rejoinder by the authors journal, November 1999

Parallel distributed computing using Python journal, September 2011

The ensemble Kalman filter for combined state and parameter estimation journal, June 2009

A universal formula for stabilization with bounded controls journal, June 1991

Data Driven Economic Model Predictive Control journal, April 2018

Recurrent neural network for solving model predictive control problem in application of four-tank benchmark journal, May 2016

Model Predictive Control of Unknown Nonlinear Dynamical Systems Based on Recurrent Neural Networks journal, August 2012

An Ensemble Kalman Filter for Systems Governed by Differential Algebraic Equations (DAEs) journal, January 2012

On the implementation of an interior-point filter line-search algorithm for large-scale nonlinear programming journal, April 2005

A framework for parallel and distributed training of neural networks journal, July 2017

Constrained Nonlinear State Estimation Using Ensemble Kalman Filters
journal, March 2010

A parallel computing platform for training large scale neural networks
conference, October 2013

The Ensemble Kalman Filter: theoretical formulation and practical implementation
journal, November 2003

Economic model predictive control of nonlinear process systems using empirical models
journal, November 2014

David Draper and E. I. George, and a rejoinder by the authors
journal, November 1999

Parallel distributed computing using Python
journal, September 2011

The ensemble Kalman filter for combined state and parameter estimation
journal, June 2009

A universal formula for stabilization with bounded controls
journal, June 1991

Data Driven Economic Model Predictive Control
journal, April 2018

Recurrent neural network for solving model predictive control problem in application of four-tank benchmark
journal, May 2016

Model Predictive Control of Unknown Nonlinear Dynamical Systems Based on Recurrent Neural Networks
journal, August 2012

An Ensemble Kalman Filter for Systems Governed by Differential Algebraic Equations (DAEs)
journal, January 2012

On the implementation of an interior-point filter line-search algorithm for large-scale nonlinear programming
journal, April 2005

A framework for parallel and distributed training of neural networks
journal, July 2017